Integrate MongoDB with n8n for AI Workflows

You can integrate MongoDB with n8n to build automations and agentic workflows using their no-code visual interface. This page provides an overview of the integration and describes the different types of nodes that you can use in your workflows.

Get Started with n8n

To use the MongoDB nodes in n8n, complete the following steps:

Install n8n.

To install n8n, see the n8n documentation or run the following command by using npm:

npx n8n

Configure your MongoDB credentials in n8n.

All MongoDB n8n nodes require MongoDB credentials. To learn how to add your credentials, see MongoDB Credentials in the n8n documentation.

Create a vector search index.

If you plan to use the MongoDB Atlas Vector Store Node, create a MongoDB Vector Search index on your collection. To learn how to create the index by using the MongoDB node in n8n, see Create the Vector Search Index in n8n.

To learn how to build a basic AI agent using n8n and MongoDB, see Build an AI Agent with MongoDB and n8n.

MongoDB Node

The MongoDB node allows you to automate work in MongoDB and integrate MongoDB with other nodes in your n8n workflows.

MongoDB Node Usage

Use the MongoDB node in any of your custom n8n workflows.

MongoDB Node Operations

The MongoDB node supports the following operations:

Category	Operation	Description
Document Operations	Aggregate Documents	Perform aggregation operations to process and transform data using MongoDB aggregation pipelines. To learn more, see `db.collection.aggregate()`.
	Find Documents	Query and retrieve documents from your MongoDB collections with flexible filtering options. To learn more, see `db.collection.find()`.
	Insert Documents	Add new documents to your MongoDB collections. To learn more, see `db.collection.insertMany()`.
	Update Documents	Modify existing documents in your collections. To learn more, see `db.collection.updateOne()`.
	Delete Documents	Remove documents from your collections. To learn more, see `db.collection.deleteMany()`.
	Find and Replace Documents	Search for documents and replace them with new content. To learn more, see `db.collection.findOneAndReplace()`.
	Find and Update Documents	Search for documents and update specific fields. To learn more, see `db.collection.findOneAndUpdate()`.
Search Index Operations	Create Search Indexes	Create new search and vector search indexes on your collections. To learn more, see `db.collection.createSearchIndex()`.
	List Search Indexes	Retrieve information about existing search indexes. To learn more, see `$listSearchIndexes`.
	Update Search Indexes	Modify existing search index configurations. To learn more, see `db.collection.updateSearchIndex()`.
	Drop Search Indexes	Remove search indexes that are no longer needed. To learn more, see `db.collection.dropSearchIndex()`.

Tip

To learn more, see n8n MongoDB node documentation

MongoDB Atlas Vector Store Node

The MongoDB Atlas Vector Store node enables you to use MongoDB Vector Search in your agentic workflows.

To use this node, you must have a MongoDB Vector Search index on your collection. To learn how to create the index in n8n, see Create the Vector Search Index in n8n.

Vector Store Node Usage

Use the MongoDB Vector Store node in the following workflow patterns:

Connect directly to an AI agent as a tool to perform agentic RAG.

AI Agent (tools connector) → MongoDB Vector Store

For a tutorial, see Build an AI Agent with MongoDB and n8n.

To learn more about AI agents in n8n, see AI agent node.

Use the MongoDB Atlas Vector Store as a regular node to insert or retrieve documents in your custom workflows:

Trigger → MongoDB Vector Store (Insert/Get) → Next Node

To learn more, see Nodes.

Use the node as a retriever in a question-answering chain:

Question and Answer Chain → Vector Store Retriever → MongoDB Vector Store

To learn more about Q&A in n8n, see Question and Answer Chain node.

Use the node as a question-answering tool for an AI agent:

AI Agent → Vector Store Question Answer Tool → MongoDB Vector Store

To learn more about the question-answering tool in n8n, see Vector Store Question Answer Tool node.

Vector Store Operation Modes

The MongoDB Vector Store node supports the following operation modes. The retrieve document mode is only available in certain workflow patterns.

Operation Mode	Description
Get Many	Retrieve multiple documents using similarity search based on a prompt. Returns documents with similarity scores. To learn more, see Get Many parameters.
Insert Documents	Add new documents with vector embeddings to your collection. To learn more, see Insert Documents parameters.
Retrieve Documents (As Vector Store for Chain/Tool)	Only available when you use the node as a retriever or a tool. Must be connected to a retriever node or root node. To learn more, see Retrieve Documents (As Vector Store for Chain/Tool) parameters.
Retrieve Documents (As Tool for AI Agent)	Only available when you use the node as a tool for an AI agent. The agent uses this vector store when the name and description are relevant to the prompt. To learn more, see Retrieve Documents (As Tool for AI Agent) parameters.

Vector Store Node Parameters

Category	Setting	Operation Mode	Description
Common Parameters	MongoDB Collection	All	Name of the MongoDB collection to use.
	Vector Index Name	All	Name of the Vector Search index in your MongoDB collection.
	Embedding Field	All	Field name in your documents that contains the vector embeddings.
	Metadata Field	All	Field name in your documents that contains the text metadata.
Mode-Specific Parameters	Name	Retrieve Documents (As Tool for AI Agent)	Name of the vector store tool for the AI agent.
	Description	Retrieve Documents (As Tool for AI Agent)	Explanation for the LLM about what this tool does.
	Limit	Retrieve Documents (As Tool for AI Agent)	Number of results to retrieve from the vector store.
Additional Options	Metadata Filter	Get Many, Retrieve Documents (As Tool for AI Agent), Retrieve Documents (As Vector Store for Chain/Tool)	Filter results based on metadata criteria.
	Rerank Results	Get Many, Retrieve Documents (As Tool for AI Agent), Retrieve Documents (As Vector Store for Chain/Tool)	Enable result reranking (requires connecting a reranker node).

Tip

To learn more, see n8n MongoDB Vector Store node documentation

MongoDB Chat Memory Node

The MongoDB Chat Memory node allows you to use MongoDB as a memory store for storing chat history in your AI workflows. This enables persistent conversation context across workflow executions.

Chat Memory Node Usage

You must use the MongoDB Chat Memory node as a sub-node by adding it to the Memory section of an AI agent node. For a tutorial, see Build an AI Agent with MongoDB and n8n.

Note

If you add multiple MongoDB Chat Memory nodes to your workflow, all nodes access the same memory instance by default. For separate memory instances, use different session IDs in each memory node.

Chat Memory Node Parameters

Parameter	Description
Session ID	Method for determining how the session key is identified. You can define the session key through a connected trigger, or you can define the key manually.
Session Key	Unique identifier for the chat session.
Collection Name	Name of the collection to store chat history. MongoDB creates the collection if it doesn't exist. Defaults to `n8n_chat_histories`.
Database Name	Name of the database to store chat history. If not provided, n8n uses the database from credentials.
Context Window Length	Number of previous interactions to consider for context.

Tip

To learn more, see n8n MongoDB Chat Memory node documentation

Additional n8n Resources

To learn more about n8n, use the following resources:

Back

Get Started

Build an AI Agent